XPS Study of Ion Irradiated and Unirradiated UO2 Thin Films

Aug 4, 2016 - XPS Study of Ion Irradiated and Unirradiated UO2 Thin Films. Yury A. Teterin,. †,‡. Aleksej J. Popel,*,§. Konstantin I. Maslakov,. ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/IC

XPS Study of Ion Irradiated and Unirradiated UO2 Thin Films Yury A. Teterin,†,‡ Aleksej J. Popel,*,§ Konstantin I. Maslakov,† Anton Yu. Teterin,‡ Kirill E. Ivanov,‡ Stepan N. Kalmykov,†,‡ Ross Springell,⊥ Thomas B. Scott,⊥ and Ian Farnan§ †

Chemistry Department, Lomonosov Moscow State University, Moscow 119991, Russia NRC “Kurchatov Institute”, Moscow, 123182, Russia § Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, United Kingdom ⊥ Interface Analysis Centre, School of Physics, University of Bristol, Bristol BS8 1TL, United Kingdom ‡

ABSTRACT: XPS determination of the oxygen coefficient kO = 2 + x and ionic (U4+, U5+, and U6+) composition of oxides UO2+x formed on the surfaces of differently oriented (hkl) planes of thin UO2 films on LSAT (Al10La3O51Sr14Ta7) and YSZ (yttria-stabilized zirconia) substrates was performed. The U 4f and O 1s core− electron peak intensities as well as the U 5f relative intensity before and after the 129 Xe23+ and 238U31+ irradiations were employed. It was found that the presence of uranium dioxide film in air results in formation of oxide UO2+x on the surface with mean oxygen coefficients kO in the range 2.07−2.11 on LSAT and 2.17−2.23 on YSZ substrates. These oxygen coefficients depend on the substrate and weakly on the crystallographic orientation. On the basis of the spectral parameters it was established that uranium dioxide films AP2,3 on the LSAT substrates have the smallest kO values, and from the XRD and EBSD results it follows that these samples have a regular monocrystalline structure. The XRD and EBSD results indicate that samples AP5−7 on the YSZ substrates have monocrystalline structure; however, they have the highest kO values. The observed difference in the kO values was probably caused by the different nature of the substrates: the YSZ substrates provide 6.4% compressive strain, whereas (001) LSAT substrates result only in 0.03% tensile strain in the UO2 films. 129Xe23+ irradiation (92 MeV, 4.8 × 1015 ions/cm2) of uranium dioxide films on the LSAT substrates was shown to destroy both long-range ordering and uranium close environment, which results in an increase of uranium oxidation state and regrouping of oxygen ions in uranium close environment. 238U31+ (110 MeV, 5 × 1010, 5 × 1011, 5 × 1012 ions/cm2) irradiations of uranium dioxide films on the YSZ substrates were shown to form the lattice damage only with partial destruction of the long-range ordering.



INTRODUCTION Uranium dioxide, UO2, is the main form of nuclear fuel used in the present generation of nuclear reactors. The knowledge and understanding of its in-reactor behavior and its stability under subsequent storage and disposal conditions are of great technological importance.1,2 The heat generated at nuclear power plants comes primarily from the slowing down of fission products with energies in the range from 70 to 100 MeV. As a result, heat and radiation damage are produced inside the fuel pellets.3−5 Fresh fuel has close to stoichiometric (UO2.001) composition. However, under in-reactor irradiation the fuel might develop an increased degree of nonstoichiometry.1 Solubility and dissolution of uranium oxide in aqueous environment strongly depends on uranium valence state, as U(VI) is more soluble than U(IV) by many orders of magnitude.6 Hence, the degree of nonstoichiometry in spent nuclear fuel has an important effect on its solubility and corrosion rate7 which governs the release rate of the majority of radionuclides.8 This work considers the explicit effect of radiation damage by fission fragments on nonstoichiometry in spent nuclear fuel and outlines a methodology developed for determining the degree of nonstoichiometry in UO2+x. For this purpose, thin films of © 2016 American Chemical Society

UO2 on LSAT (lanthanum strontium aluminum tantalum oxide) and YSZ (yttria-stabilized zirconia) substrates were produced and irradiated with Xe and U ions, respectively. The irradiated and unirradiated films were analyzed by EDX (energy-dispersive X-ray spectroscopy), XRD (X-ray diffraction), EBSD (electron backscatter diffraction), SEM (scanning electron microscopy), and XPS technique. The obtained results were compared. Previous papers considered mechanisms of xenon transfer and its interaction with uranium in UO2+x (ref 9) theoretically, as well as the influence of defects10,11 and pressure12,13 on the ionic composition of these oxides. Adsorption energies of water on differently oriented uranium dioxide planes were calculated.14−16 The work on the study of the electronic structure of uranium and its alloys17−21 and oxides UO2+x (x ≤ 0 and x ≥ 0) employed theoretical calculation results,22−34 photoelectron spectroscopy (PES),35−41 and X-ray photoelectron spectroscopy (XPS) widely.38,42−55 These techniques were used to study films on various substrates.38−40,56−58 Received: May 13, 2016 Published: August 4, 2016 8059

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

of single-crystalline films used the technique developed on the basis of the dependence of the U 5f peak relative intensity I5f (rel. units) on the oxygen coefficient kO.44,45,49

Determination of the uranium oxidation state and sample’s ionic composition employs the U 4f doublet split due to the spin−orbit interaction by ΔEsl (U 4f) = 10.8 eV.17,50,59 The binding energy (BE) Eb(U 4f7/2) of the U 4f7/2 electrons in uranium and oxides grows as ∼377 (metallic U), ∼380 (U4+), ∼381 (U5+), and ∼382 (U6+).17,50,55,56,59−62 Special attention was paid to the study of the mechanisms of structure formation, which leads to widening of the main peaks and the appearance of extra structure in the spectra.50,59,63 The XPS spectra of some oxides exhibit typical shake-up satellites of about ∼25% intensity of the basic peaks.41,51,60,64 The calculated spectroscopic factors fAnf reflecting the fractions of the basic peak XPS intensities with the deduction of shake-up satellite intensities for the U 4f and U 5f peaks are f U 4f = 0.83 and f U 5f = 0.86.65,66 Shake-up satellites are located from the basic peaks toward the higher BE by ΔEsat(U 4f): ∼7 (U4+), ∼8 (U5+), ∼4, and ∼10 eV (U6+).60 Since the U 4f BE on moving from UO2 to UO3 changes by ΔEb ≈ 2 eV, one can reliably determine uranium oxidation states for individual uranium oxides on the basis of the U 4f7/2 BE and positions (ΔEsat) and relative intensities Isat (%) of the shake-up satellites.56,60 The U 4f XPS structure is best resolved for the crystalline oxides. For complex amorphous oxides UO2+x it is often difficult to segregate unambiguously the U 4f XPS peaks containing shake-up satellites into separate components for reliable quantitative information on uranium oxidation state and ionic composition of the studied oxide. Despite this, the XPS studies of uranium oxides by authors of refs 38, 47, 50, 60, and 62 employed the U 4f XPS peak decomposition with shake-up satellite parameters in mind. It is known that the traditional method of determination of the oxygen coefficient kO = 2 + x for UO2+x based on the U 4f/ O 1s XPS intensity ratio (with photoionization cross sections or experimental sensitivity factors in mind) does not give satisfactory results (see ref 49). It is due to the fact that the O 1s intensity grows due to the presence of complex oxides UO2+x as well as impurity oxides containing excess oxygen on the sample surface. The valence electron BE range (from 0 to ∼35 eV) in XPS of oxides UO2+x changes significantly as the oxygen coefficient grows.45 Thus, on moving from UO2 to γ-UO3 in the outer valence molecular orbitals (OVMO) with BE range (from 0 to ∼15 eV) a sharp U 5f peak disappears and in the inner valence molecular orbitals (IVMO) with BE range (from ∼15 to ∼35 eV) instead of a single atomic U 6p3/2 peak two components appear.42,61 Such a splitting is due to the IVMO formation in γUO3. The IVMO formation in uranium oxides due to interaction of the U 6p and O 2s atomic orbitals (AO) also results in formation of structure in other X-ray (emission, conversion, Auger) spectra of uranium oxides.59,67−69 These structure parameters correlate with the uranium−oxygen interatomic distances in axial and equatorial directions in uranyl compounds and serve for quantitative determination of these compounds.50,59 The peak of the U 5f electrons weakly participating in the chemical bond in UO2+x is observed in the photoelectron70 and X-ray photoelectron52−54,68,69 spectra around 0 BE. Spectra of compounds containing U6+ do not exhibit this peak.61 Therefore, the present work for determination of the oxygen coefficient kO = 2 + x, uranium oxidation state and ionic composition k (%) of the studied oxides UO2+x on the surface



EXPERIMENTAL SECTION

Samples: Thin Films Production. Thin films of epitaxial UO2 were produced by reactive sputtering onto LSAT and YSZ substrates with three different crystallographic orientations: (001), (110), and (111).71 A dedicated dc magnetron sputtering facility with UHV base pressure (10−9 mbar) was employed to grow the films. A depleted uranium metal target was used as a source of uranium. It was kept at a power of 50 W by controlled direct current of 0.11−0.14 A and the corresponding voltage of 350−450 V, giving a growth rate of 0.9−1.1 Å/s for films on the LSAT substrates, and by controlled direct current at an average value of 0.15 A and the corresponding voltage of 330 V, giving a growth rate of about 1.5 Å/s for films on the YSZ substrates. Argon was used as the sputtering gas at a pAr in the range of 7−8 × 10−3 mbar. Oxygen was used as the reactive gas at a pO2 in the range 3.4 to 4.4 × 10−5 mbar for films on the LSAT substrates and at a pO2 of 2 × 10−5 mbar for films on the YSZ substrates, except for sample OB6 for which a pO2 was 3 × 10−6 mbar. The LSAT and YSZ substrates were kept at a temperature close to 750 and 600 °C, respectively. The substrates were one side polished single-crystal LSAT or YSZ with dimensions of 10 × 10 × 0.5 mm supplied by MTI Corp., USA. LSAT has a cubic perovskite structure with aLSAT = 3.868 Å,72 and UO2 has a cubic fluorite structure with auo2 = 5.469 Å, both at room temperature.71 This results in the epitaxial relationship in which the (001) plane of UO2 is rotated by 45° in relation to the (001) plane of LSAT so that the (110) plane of UO2, with a d spacing of auo2/√2 = 3.867 Å, fits the LSAT (001) plane, with a d spacing of 3.868 Å (= aLSAT), as described by Bao et al.71 for a UO2 film on a LaAlO3 substrate. This causes the UO2 lattice to be only at a slight tension of +0.03% with respect to the substrate in-plane spacing. This 45° rotation epitaxial relationship only holds between the LSAT and the UO2 (001) planes. YSZ has a cubic fluorite structure with aYSZ = 5.139 Å.73 This results in the plane to plane epitaxial match in which the plane of UO2 is put at compression of −6.4% by the plane of YSZ. This plane to plane epitaxial relationship holds for the (001), (110), and (111) plane orientations. Table 1 summarizes the produced samples. Sample pairs AP1/OB1−AP4/OB4 were produced by cutting one sample into two halves using a diamond saw for various studies. It was identified by means of EDX and XPS analyses that the films of uranium dioxide contain Nb. On the basis of the results from EDX, XRD, EBSD, SEM, and XPS measurements, it is suggested that Nb is present in the form of Nb2O5 and is located in particulates, which precipitated onto the substrates during growth of the films. The particulates can be seen in SEM images (not shown) obtained at

Table 1. Summary of the Produced Thin Film UO2 Samples sample

LSAT (AP1-AP4) and YSZ (AP5-AP7, OB5-OB7) substrate crystallographic orientation (hkl)

UO2 film orientation (hkl)

UO2 film thickness (nm)

AP1 AP2 AP3 AP4 AP5 OB5 AP6 OB6 AP7 OB7

(111) (001) (001) (110) (001) (001) (110) (110) (111) (111)

(210)a (001) (001) (111)a (001) (001) (110) (110) (111) (111)

110 140 120 140 90 150 150 150 150 150

a

8060

Preferred crystallographic orientation of the as-produced samples. DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

typical fission fragments.4,5 The flux was kept at around 1.3 × 1010 ions/(cm2 s), which caused heating of the samples to a temperature not exceeding 150 °C. The samples were allowed to cool down to ambient temperature (around 19 °C) before the beamline was brought to atmospheric pressure using nitrogen gas to minimize surface oxidation of the samples. UO2 films on the YSZ substrates were irradiated with 238U31+ ions of 110 MeV energy to fluences of 5 × 1010 (OB5), 5 × 1011 (OB6), and 5 × 1012 (OB7) ions/cm2 to induce radiation damage. The flux was kept at around 1 × 108 ions/(cm2 s). The irradiation was conducted at an ambient temperature of 16−17 °C. No heating of the samples was observed. The beamline base vacuum was 6 × 10−7 mbar during irradiation. According to the SRIM-2012.03 software,74 the nuclear and electronic stopping, dE/x, for 92 MeV 129Xe23+ ions in UO2 is 0.26 and 24.6 keV/nm, respectively, and the projected range is 6.5 μm and for 110 MeV 238U31+ ions is 0.96 and 27.4 keV/nm, respectively, and the projected range is 6.7 μm. A theoretical UO2 density of 10.96 g/ cm3 (ref 5) was assumed in the SRIM calculation. The SRIM results indicate that the 129Xe23+ and 238U31+ ions completely penetrate the UO2 thin films (150 nm max), and the electronic stopping regime dominates the dissipation of ion energy throughout the entire film. X-ray Photoelectron Measurements. XPS spectra of UO2+x were recorded on a Kratos Axis Ultra DLD spectrometer using monochromatic Al Kα radiation (hν = 1486.6 eV) at 150 W X-ray gun power under 10−9 mbar at room temperature (Figure 2). The analyzed

different angles. They have sizes down to 30 nm and are densely populated. Niobium concentration was determined by the EDX technique, with the point-analysis spot size of 1−2 μm in diameter, at different locations on the surface of the films (∼3.5 wt %), which agrees with the concentration values obtained from XPS (∼5 wt %). Since the analysis spot in XPS is an ellipse with minor and major axes of 300 and 700 μm, respectively, it was not possible to find an area free of niobium oxide. That is why Nb2O5 lines were observed in XPS spectra from the studied samples which had a relatively small width Γ(Nb 3d5/2) = 1.2 eV and Eb(Nb 3d5/2) = 206.9 eV. This agrees with the fact that the oxidation state of Nb in the samples is only Nb5+. The work by Fu et al.19 also shows that regions of UO2 and Nb2O5 are formed during the oxidation of uranium−niobium alloy with oxygen. Niobium oxide is not observed in XRD scans of the samples, possibly due to its low concentration in the films. This observation is consistent with the results obtained in the work by Strehle et al.,73 where the codeposition of U and Nd was performed onto YSZ substrates. The XRD scans did not exhibit any difference between pure UO2 and UxNdyOz samples, which can be related to the absence of the crystal structure involving neodymium. Crystallographic orientations for the UO2 films were determined by means of XRD (θ−2θ scans with φ rotation) and EBSD techniques (not shown). In addition, UO2 films on the (001) YSZ substrates, produced under similar conditions, were thoroughly characterized by Strehle et al.,73 and it was shown that these films are single crystals. Since the epitaxial relationship and lattice mismatch for the UO2 films on the (111) and (110) YSZ substrates are the same as for the UO2 films on the (001) YSZ substrate and based on the obtained XRD and EBSD results, it is possible to suggest that these films are also single crystals. On the basis of the expected epitaxial relationship for UO2 films on the (001) LSAT substrates and the obtained XRD and EBSD results, we are inclined to suggest that the films on the (001) LSAT substrates can be considered as single crystals. As an example, Figure 1

Figure 1. Triangular inverse pole figure for sample AP3 obtained from the EBSD study. Figure 2. Survey XPS scan from a UO2 film (sample AP3).

illustrates a result of the EBSD study for sample AP3, which confirms formation mainly of a single crystal with the surface orientation (001). Sample AP2 provides similar results. Uranium dioxide films on the (111) and (110) LSAT substrates are described as preferentially oriented. Film thickness for samples AP1−AP4 was measured using transverse SEM on a cross section of the sample. Focused ion beam (FIB)-SEM was used to measure film thickness for samples OB6 and OB7 after they have been irradiated. Secondary ion mass spectrometry (SIMS) was performed on sample OB6 (after it has been irradiated) to verify the thickness measured with the FIB-SEM method and on sample AP6 to deduce its thickness, as it was not possible to resolve the UO2 film in the FIB-SEM study, possibly, due to a low electrical conductivity of the film. Film thickness for samples AP5, OB5, and AP7 was estimated based on the growth rate calculated from the measured film thicknesses for samples AP6, OB6, and OB7 and the corresponding deposition times. Sample Irradiation. Sample irradiation was performed on the IRRSUD beamline at the GANIL accelerator, Caen, France. UO2 films on the LSAT substrates (OB1−4) were irradiated with 129Xe23+ ions of 92 MeV energy to a fluence of 4.8 × 1015 ions/cm2 to simulate the damage produced by fission fragments in nuclear fuel. The energy and mass of the ions used for the irradiation are representative of the

area was an ellipse with 300 and 700 μm minor and major axes, respectively. Binding energy scale of the spectrometer was preliminarily calibrated by the position of the peaks of Au 4f7/2 (83.96 eV) and Cu 2p3/2 (932.62 eV) core levels for pure gold and copper metals. The spectra were acquired in the constant analyzer energy mode using a pass energy of 20 eV and a step size of 0.05 eV. The equipment resolution measured as the full width on the halfmaximum (fwhm) of the Au 4f7/2 peak was less than 0.65 eV. The binding energies (BE) were measured relative to the BE of the C 1s electrons from hydrocarbons adsorbed on the sample surface that was accepted to be equal to 285.0 eV. The fwhms are given relative to that of the C 1s XPS peak from hydrocarbon on the sample surface being 1.3 eV.50 The error in the determination of the BE and the peak width did not exceed ±0.05 eV, and the error of the relative peak intensity was ±5%. The inelastically scattered electrons-related background was subtracted with the Shirley method.75 40 + Ar etching of 2 × 2 mm2 sample area was conducted at the accelerating voltage of 2 kV and beam current of 50 μA at 2 × 10−9 mbar and room temperature. The etching rate was 7.1 nm/min for SiO2 under these conditions. Thus, UO2 film on the YSZ substrate 8061

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

Table 2. Surface Elemental Composition of the Epitaxial UO2+xa Thin Films, Peak Intensity of the U 5f Electrons,b Oxygen Coefficient kOc in UO2+x Oxide, Composition of Uranium Ions k (%)d of Unirradiated (AP1-7) and 129Xe23+ (OB1-4), and 238 31+ U (OB5-7) Irradiated Thin Films of Uranium Dioxide k no. 1 2 3 4 5 6 7

sample (hkl) e

AP1 (210) OB1 AP2 (001) OB2 AP3 (001) OB3 AP4 (111)e OB4 AP5 (001) OB5 (001) AP6 (110) OB6 (110) AP7 (111) OB7 (111) AP7(Ar+)f

elemental composition UO2+x

IU 5f (±0.001)

kO in UO2+x (±0.01)

U4+

UO2.8 UO47 UO3.5 UO21 UO3.6 UO15 UO3.9 UO15 UO3.4 UO3.4 UO3.2 UO3.2 UO3.6 UO3.4 UO1.98

0.025 0.013 0.029 0.025 0.028 0.023 0.027 0.021 0.021 0.022 0.017 0.021 0.019 0.020 0.025

2.11 2.31 2.07 2.11 2.08 2.14 2.10 2.17 2.17 2.15 2.23 2.17 2.20 2.18 2.11

44 (45) 0.3 58 (60) 44 52 (58) 33 48 (54) 25 25 (47) 32 (45) 13 (37) 25 (44) 18 (44) 23 (48) 42

U5+ 45 68 35 45 39 52 42 57 57 53 64 57 61 59 47

(40) (36) (38) (41) (48) (47) (55) (48) (48) (45)

U6+ 11 (15) 31 7 (4) 11 9 (4) 14 10 (5) 17 17 (5) 15 (8) 23 (8) 17 (8) 21 (8) 18 (7) 11

a Elemental composition obtained on the basis of the core line U 4f7/2 and O 1s intensities of uranium dioxide and atomic photoionization cross sections σ: 2.81 (O 1s), 36.0 (U 4f7/2). bPeak intensity of the U 5f electrons measured as a ratio: IU 5f = I(U 5f)/I(U 4f7/2) without taking into account intensities of the shake-up satellites. cOxygen coefficient kO = 2+x in UO2+x oxide found from eq 1. dComposition of uranium ions k (%) found by using eqs 4−6; the values obtained based on dividing the U 4f7/2 peak into components and intensities of the shake-up satellites are shown in the parentheses. ePreferred crystallographic orientation. fSample AP7 after the 180 s Ar+ treatment.

(AP7) was Ar+ etched at 2 keV and 2 × 10−9 mbar. The flux was maintained at about 3.1 × 1014 ions/(cm2 s). The quantitative elemental analysis was performed for several nanometer-deep layers of the studied samples. It was based on the fact that the spectral intensity is proportional to the number of certain atoms in the studied sample. The following ratio was used: ni/nj = (Si/ Sj)(kj/ki), where ni/nj is the relative concentration of the studied atoms, Si/Sj is the relative core−shell spectral intensity, and kj/ki is the relative experimental sensitivity coefficient. The following coefficients relative to C 1s were used: 1.00 (C 1s); 2.81 (O 1s); 10.51 (Nb 3d); 36.0 (U 4f7/2; see ref 76). Determination of the Oxygen Coefficient kO = 2 + x and Ionic Composition of Oxides UO2+x. Once uranium oxide UO2+x contains uranium ions of different oxidation states, the U 4f XPS spectrum is expected to consist of several peaks at different BEs. Usually these peaks are superimposed because of low equipment resolution. This widens and distorts the shape of the main peak, which complicates its decomposition into components. The O 1s intensity in this case is usually higher due to oxygen-containing impurities on the surface. This increases the error in the oxygen coefficient kO = 2 + x determination, which is found from the U 4f/O 1s intensity ratio with sensitivity coefficients in mind. In this case we can only yield a qualitative interpretation of the results (Table 2, column 3). Determination of uranium ions composition on the basis of intensities of U 4f-electron lines was carried out based on the known parameters of the spectra of uranium oxides. Both the primary and the satellite peaks for U 4f were used in the fitting procedure. The binding energy of the U 4f7/2 electrons was taken as ∼380 (U4+), ∼381 (U5+), and ∼382 eV (U6+) and shake-up satellites located from the basic peaks toward the higher BE by ∼7 (U4+), ∼8 (U5+), and ∼4 eV and ∼10 eV (U6+) were used, as described in the Introduction. Although it is known that the fwhm, Γ(U 4f7/2), of the U 4f7/2 line decreases as ∼1.5 (U4+), ∼1.4 (U5+), and ∼1.2 eV (U6+), related to multiplet splitting and decrease of unpaired U 5f electrons from 2 to 0, the fwhm value of 1.4 eV was taken for fitting the U4+, U5+, and U6+ curves to simplify the fitting. The curve shapes were approximated by a mixed Gaussian (∼80%) and Lorentzian (∼20%) function to get the best fit to the experimental curve. In cases when some peaks were absent, the initial % composition was determined based on the method of U 5f/U 4f7/2-electron intensities (our suggested method), followed by an

iterative fitting procedure to obtain a final fit. The results obtained by the spectra fitting procedure are shown in parentheses in Table 2 in the fourth column. The oxygen coefficient kO = 2 + x and ionic composition of UO2+x can be also determined on the basis of the intensity of the peak of the U 5f electrons not participating in the chemical bonding. The present work physically grounded this technique and considered it in more details than before.44,45,49 This technique considers oxygen ions directly bound with uranium ions. Adsorbed oxygen ions not participating in uranium−oxygen bonding do not affect the U 5f intensity; however, they affect strongly the uranium/oxygen ratio on the surface, which practically does not allow the traditional XPS quantitative analysis to be used. The U 5f electrons weakly participating in chemical bonding in uranium oxides are strongly localized and observed as a sharp peak at the lower BE side from the outer valence band (see Figure 3). The XPS from γ-UO3 not containing the U 5f electrons does not exhibit this peak.61 Since the U 5f BE is about ∼3 eV lower than that of the low-energy OVMO band edge (“quasi-gap”), one can suggest that the U 5f intensity must decrease discretely as uranium oxidation state grows from U4+ to U5+ and U6+, since the U 5f electrons transit from the localized state to the outer valence band. The U 5f intensity must first decrease twice and then vanish (Figure 3). Indeed, this was observed in the XPS of neptunium compounds, as neptunium oxidation state increased from Np(5f2)5+ to Np(5f1)6+ (ref 50). Therefore, the U 5f intensity can be used as a quantitative parameter of the number of U 5f electrons involved in chemical bonding in uranium oxides. The relative U 5f intensity I1 (rel. units) determined as the U 5f/U 4f7/2 intensity ratio without the shake-up satellites can be presented as dependence on the oxygen coefficient kO = 2 + x (Figure 4). For synthetic and natural oxides it is44,45

I1 = 5.366k O−7.173

(1)

which is in good agreement with the dependence of the magnetic susceptibility on the oxygen coefficient.50,77 This agreement is not unexpected since the U 5f intensity is proportional to the number of the U 5f electrons responsible for the magnetic properties. The value I1(2) = 0.0383 for UO2 was found by the extrapolation of dependence (I1) at kO → 2 (Figure 4). It differs from the value 0.0372 found from eq 1. The value I1(3) = 0.002 is not zero at kO → 3 8062

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry I3 = I2 − 2(I2 − I1) = 2I1 − I2

(3)

This dependence shows (see Figure 4) that as kO grows from 2 to 2.35 the U4+ ions in UO2+x vanish. This agrees with the XRD data showing that in synthetic oxides at kO → 2.38 the UO2 structure containing the U4+ ions vanishes.78 In this approximation, eqs 1−3 and Figure 4 enable one to determine the oxygen coefficient kO and ionic composition in UO2+x. Indeed, if the value I1 is known from the experiment, the values I2 and I3 can be found from eqs 2 and 3 and fractions ν (Un+) of uranium ions of different oxidation states can be found from

ν1(U 4 +) = I3/0.0383

(4)

ν2(U5 +) = 2(I2 − I1)/0.0383

(5)

ν3(U 6 +) = (0.0383 − I2)/0.0383

(6)

In this case we can obtain repetitive data agreeing with other studies results.50 The data obtained using eqs 1−6 are given in Table 2. In such an approximation the U5+ state was considered as a spectroscopic state for dependence of I1 where the oxygen coefficients were obtained using chemical methods.45 The U5+ ions were suggested to appear in UO2 (CaF2) lattice due to an increase of some of U−O interatomic distances without changes in the stoichiometric composition of UO2. Physically it means that in the U(IV)O2 phase another phase U(V)O2 forms without changes in oxygen content. It leads to disappearance of CaF2 lattice in UO2+x as the oxygen coefficient grows, which agrees with the XRD data.78 Therefore, in order to deduce the oxygen coefficient kO from the ion fractions νi (4,5,6) one has to multiply by 2 (the number of oxygens in UO2) the sum of fractions [ν1(U4+) + ν2(U5+)] and to add the fraction ν3(U6+) multiplied by 3 (the number of oxygens in UO3) (see Table 2). The error in the fractions of UO2+x ionic composition has to be taken into account.

Figure 3. Valence XPS scan from a UO2 film (sample AP3). Vertical bars show the calculated spectrum for the UO812− (Oh) cluster reflecting uranium close environment in UO2 (ref 42).



RESULTS AND DISCUSSION Determination of the uranium oxidation state and UO2+x ionic composition, as mentioned above, employs both the traditional XPS parameters (BEs and peak intensities) and the structure parameters of the core and valence spectra such as U 5f relative intensity, OVMO (IVMO) − core level BE differences; spin− orbit splitting ΔEsl (eV) and multiplet splitting ΔEms (eV), dynamic effect-related structure parameters, and relative positions of core-level shake-up satellites ΔEsat (eV).50 These XPS parameters allow getting information on uranium physical and chemical properties in the studied samples. The surface treatment with Ar+ ions was not used during the XPS study of the films in this work, as it is known that Ar+ etching can change the ionic composition of the surface. Since the data from the surface etched by Ar+ can be useful in discussing the results of the study of the radiation damage of the uranium dioxide films, for sample AP7 the effect of Ar+ etching on the surface composition was studied. It was found that after the 20 s etching the C 1s intensity became ∼10 times lower and the O 1s peak became single with Γ(O 1s) = 1.2 eV. The U 4f spectrum was observed as a spin−orbit split (ΔEsl = 10.8 eV) of 1.8 eV wide doublet with symmetrical peaks. Intensity of the shake-up satellites (Isat = Is/Io), equal to the ratio of the satellite area (Isat) to the area of the basic peak (Io), was observed to have 30% intensity at ΔEsat1 = 6.9 eV. This structure is typical for UO2 (ref 50). After the 60−120−180 s etching and staying for a while in the spectrometer chamber (“annealing”) the XPS structure did not change significantly. After the 180 s etching the U 4f7/2 XPS exhibited a weak shoulder at the lower BE side attributed to metallic U. For the 180 s etching the oxygen coefficient kO was 1.98 based on the U 4f and O 1s intensities. The same coefficient found on the basis

Figure 4. Dependence of the relative U 5f XPS intensity (I) on the oxygen coefficient (kO = 2+x) for UO2+x: I1 (eq 1); I2 (eq 2); I3 (eq 3); (o) synthetic oxides; (Δ, □) uraninites; (x) mixtures of UO2 and γUO3 (ref 45).

because it is difficult to describe the experimental curve (Figure 4) with the analytical function in eq 1. Therefore, at kO → 2 and 3 the use of dependence in eq 1 must lead to an increase of the error. With I1(2) = 0.0383 in mind, a dependence of I2 reflecting the expected change of the U 5f intensity on the oxygen coefficient kO for the mixtures of UO2 and UO3 was built as

I2 = − 0.0383k O + 0.1149

(2)

This dependence is in satisfactory agreement with the corresponding experimental data on UO2 and UO3 mixtures (Figure 4). The decrease of I2 as kO grows is due to the increase of concentration of the U6+ ions that do not contribute to the U 5f intensity. The difference ΔI = I2−I1 characterizes the decrease of the U4+ concentration in complex oxides UO2+x as kO grows. Having suggested that synthetic oxides contain only uranium ions of formal oxidation states U4+, U5+, and U6+ with electronic configurations U 5f2, U 5f1, and U 5f0, respectively, one can find a dependence of decrease of intensity I3 due to the U4+ ions as kO grows 8063

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

Table 3. Electron Binding Energy Eba (eV), Line Width Γb (eV) in the Parentheses, Satellite Positions ΔEsatc (eV) of Unirradiated (AP1-4 on LSAT and AP5-7 on YSZ), and 129Xe23+ (OB1-4) and 238U31+ (OB5-7) Irradiated Thin Films of Uranium Dioxide no. 1 2 3 4 5 6 7

sample (hkl) d

AP1 (210) OB1 AP2 (001) OB2 AP3 (001) OB3 AP4 (111)d OB4 AP5 (001) OB5 (001) AP6 (110) OB6 (110) AP7 (111) OB7 (111) AP7 (Ar+)e

U 5f

U 4f7/2 U4+

1.2(1.2) 1.3(1.2) 1.3(1.2) 1.2(1.2) 1.3(1.1) 1.0(1.2) 1.2(1.2) 1.3(1.1) 1.6(1.0) 1.5(1.1) 1.6(1.1) 1.5(1.0) 1.5(1.1) 1.4(1.2) 1.2(1.6)

379.7(1.4), 6.8(3.0) 379.7(1.4), 6.8(2.0) 379.9(1.4), 6.8(1.9) 379.7(1.4), 6.8(1.9) 380.3(1.8), 380.1(1.5), 380.2(1.6), 380.1(1.6), 380.0(1.5), 380.0(1.5), 379.8(1.6)

6.1 (2.1) 6.2(2.1) 6.0(2.5) 6.1(2.1) 6.1(2.1) 6.6(2.3)

U 4f7/2 U6+

Γ U 4f7/2

O 1s

7.9(2.1)

382.2(1.4)

7.9(2.0)

382.2(1.4)

7.9(2.1)

382.4(1.4)

8.0(2.1)

382.1(1.4)

7.5(2.1) 7.7(2.1) 7.6(2.3) 7.7(2.1) 7.7(2.1) 7.6(2.6)

383.3(1.8) 382.7(1.5) 382.9(1.6) 382.8(1.6) 382.7(1.5) 382.6(1.5)

2.1 1.6 2.3 1.9 2.2 2.0 2.3 1.8 2.5 2.5 2.4 2.3 2.4 2.4 1.6

529.8(1.1) 530.1(1.4) 529.7(1.1) 530.1(1.5) 530.1(1.0) 529.8(1.4) 529.7(1.1) 530.1(1.4) 530.1(1.1) 530.1(1.1) 530.2(1.2) 530.1(1.1) 530.1(1.1) 530.0(1.1) 530.0(1.2)

U 4f7/2 U5+ 380.9(1.4), 380.4(1.5), 380.9(1.4), 380.4(1.9) 381.1(1.4), 380.3(2.0) 380.8(1.4), 380.5(1.8) 381.4(1.8), 381.2(1.5), 381.4(1.6), 381.2(1.6), 381.2(1.5), 381.1(1.5), 381.5(1.7)

a

First peak, binding energy measured relative to the Eb(C 1s) = 285.0 eV of hydrocarbons on the sample surface. bLine width reported relative to the Γ(C 1s) = 1.3 eV in parentheses. cSecond peak, satellite energy ΔEsat relative to the basic peak. dPreferred crystallographic orientation. eSample AP7 after the 180 s Ar+ treatment.

of the U 5f intensity was 2.11. The ionic composition of the sample was found to be 42% (U4+), 47% (U5+), and 11% (U6+) (Table 2). This unexpected result agrees with the fact that uranium oxides can self-organize and form a stable lattice UO2+x containing different phases.78,79 This must be taken into account in the studies of irradiation of UO2 films with xenon and uranium ions. The survey XPS spectrum (Figure 2) provides important information on the studied sample. It consists of peaks of included elements and is typical for uranium oxide. This spectrum also contains the Auger peaks of carbon (C KLL) and oxygen (O KLL) and the XPS peak at 207.1 eV identified as the Nb 3d5/2 peak of niobium in Nb2O5 (ref 80). Niobium impurity formed during sample preparation. The Nb 3p3/2 peak of Nb2O5 was observed at Eb(Nb 3p3/2) = 362.7 eV. Despite the fact that the Nb 3p1/2 peak of Nb2O5 at Eb(Nb 3p1/2) = 378.2 eV is superimposed with the U 4f7/2 peak at Eb(U 4f7/2) ≈ 379.9 eV (Table 3), the Nb 3p1/2 intensity does not contribute significantly (within the error) to the U 4f7/2 intensity since the Nb 3p1/2 peak has many times lower intensity than the U 4f7/2 one because the U 4f7/2 photoionization cross-section is ∼10 times higher than the Nb 3p1/2 one,81 and uranium content is many times higher than niobium content. The XPS spectra are shown in Figures 2, 3, 5, 6, and 7, and the corresponding BE are given in Table 3. Valence Electron Spectra Range. The valence bands of the studied samples were observed in the BE range from 0 to ∼35 eV. They consist of the OVMO (from 0 to ∼15 eV BE) and IVMO (from ∼15 to ∼35 eV BE) ranges42 (Figure 3). Peaks in the U 6p−O 2s BE range are relatively wide and structured compared to the corresponding core U 4f and O 1s ones (Table 3, Figures 5 and 6). This contradicts the Heisenberg uncertainty ratio ΔEΔτ ≈ h/2π, where ΔE is the natural width of a level from which an electron was extracted, Δτ is the hole lifetime, and h is Planck’s constant. Since the hole lifetime (Δτ) grows as the absolute atomic level energy decreases, the lower BE XPS atomic peaks are expected to be narrower. In particular, if the O 2s level was atomic, its fwhm would be lower than that of the O 1s peak being Γ(O 1s) = 1.0

Figure 5. XPS narrow-scan O 1s from a UO2 film (sample AP3).

eV (Table 3, Figure 5), which contradicts the experimental data. Taking into account the data on UO2 (refs 42 and 45), the O 2s and U 6p levels were shown not to be atomic but the IVMO related. The calculated results of the electronic structure of the cluster UO812−, representing uranium close environment in UO2, are presented under the spectrum of valence electrons (Figure 3). These results were obtained in an approximation of the relativistic method of discrete variation (RDV).42 The valence bands of the studied samples AP1−7 and OB1− 7 are formed mostly from the U 6s,6p,6d,5f,7s,7p and O 2s,2p electrons. The OVMO structure is mostly formed from the U 6d,5f,7s,7p−O 2p interaction, while the IVMO structure is mostly formed from the U 6s,6p−O 2s interaction.42,45 The IVMO-related structure of the studied samples is a superposition of the IVMO-related structures of different uranium oxides. Although complicated, this structure provides the qualitative and quantitative information on the UO2+x physical and chemical properties50 (Figure 3). For example, the intensity of the sharp (Γ(U 5f) = 1.1 eV) single peak of the weakly bound quasiatomic U 5f electrons observed at ΔEb(U 5f) = 1.3 eV is an important parameter. The relative U 5f intensity (IU 5f) found as a ratio of the U 5f to the U 4f7/2 intensities according to eq 1 allows oxygen coefficients (kO, %) to be calculated for uranium oxides UO2+x (Table 2). 8064

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

Figure 7. XPS narrow scans of U 4f from UO2 films: (a) unirradiated surface (sample AP7); (b) after 238U31+ irradiation (110 MeV, 5 × 1012 ions/cm2, sample OB7).

Figure 6. XPS narrow scans of U 4f from UO2 films: (a) unirradiated surface (sample AP3); (b) after 129Xe23+ irradiation (92 MeV, 4.8 × 1015 ions/cm2, sample OB3).

With eqs 2−6 in mind, it allowed us to find the relative content of uranium ions (k) of different oxidation states U4+, U5+, and U6+. These data are given in Table 2 for each sample. This table also contains crystallographic (hkl) indices of the uranium dioxide films and elemental composition of UO2+x found based on the O 1s and U 4f7/2 peak intensities and the corresponding sensitivity coefficients. The ionic composition of the surface found based on the results of decomposition of the U 4f spectrum into the constituent elements (Figures 6 and 7) is presented in parentheses. The environmental influence causes formation of a complex oxide UO2+x on the surface of the single-crystal UO2 film. As a result, oxygen is observed at the surface of the films at 530.1 eV BE participating in a metal−oxygen bond (Figure 5, Table 3). The oxygen coefficient changes from 2.8 to 3.9 in the UO2+x oxide (Table 2) and exceeds possible values (from 2 to 3) for this parameter even with accounting for oxygen in Nb2O5. Therefore, it is difficult to determine the oxygen coefficient in a UO2+x oxide based on the core peak intensity of oxygen and uranium. This agrees with the previous work results.49 However, the kO values found based on the U 5f intensity are in a satisfactory agreement with the results obtained based on decomposition of the U 4f spectrum for the cases, e.g., AP2, when resolution of the U 4f spectrum is relatively good (see

values in the parentheses in Table 2). These values confirm experimentally that determination of the oxygen coefficient kO based on relative U 5f intensity for UO2+x has sufficient physical justification. This complex oxide contains three different uranium ions. Results for films AP1−4 formed on the LSAT differ from those for AP5−7 formed on the YSZ. Content of uranium ions of different oxidation states in samples AP2,3 is comparable within the measurement error (see Table 2). Samples AP2 and AP3 should have the same structure and to a higher extent are single crystalline based on the XRD data. The higher content of U4+ ions and the lower content of U5+ and U6+ in AP2 compared to AP3 can be explained by a higher (approximately by a factor of 5) carbon content on the surface of film AP2 as compared to AP3. Possibly, this carbon coating inhibits further oxidation of the film. Samples AP1 and AP4 contain 2.8 and 3.9 oxygens per one uranium atom with Eb(O 1s) ≈ 529.8 eV bonded with metal (Table 3). Despite this, the kO values for AP1 and AP4 are similar. By going from AP2,3 (UO2 (001)) to AP4 (UO2 preferential (111)) the value of kO increases. By going from samples AP1−4 to AP5−7 an increase in kO values and concentrations of U5+ and U6+ ions and a decrease in U4+ ions content is observed. Despite the difference in crystallographic orientation of the films in samples AP5−7 8065

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

can influence formation of a different number of defects during the growth of UO2 single crystals. The YSZ substrates differ from the LSAT substrates by the fact that they put the UO2 films at compression (−6.4%) due to lattice mismatch between the films and the substrates. This might lead to creation of defects in the UO2 films. The LSAT substrate in (001) orientation gives a minor mismatch of 0.03% by putting the film at tension. 238 31+ U (110 MeV, fluence 5 × 1010, 5 × 1011, and 5 × 1012 ions/cm2, penetration depth ≈ 6.7 μm) irradiation of OB5−7 films on the YSZ substrates was also performed in order to study the effect of accumulating radiation damage. The surface elemental composition of UO2+x, U 5f intensities, oxygen coefficients (kO), and ionic compositions (k, %) are given in Table 2 for each sample. 238U31+ irradiations do not cause severe damage of the films (Figure 7b). This is mainly due to low fluencies. The surface elemental composition of the UO2+x films, including carbon content, does not change within the experimental error. However, the 238U31+ irradiations cause changes in the ionic composition of the samples (Table 2). The oxygen coefficient (kO) decreases compared to that for the unirradiated counterparts. U4+ concentration grows, while U6+ concentration decreases. The error in determination of the composition of uranium ions based on decomposition of uranium peaks into components increases due to blurring of the structure in the U 4f spectra. After irradiation the composition of uranium ions on the surface of samples OB5−7 equalizes, which is observed during formation of stable (metastable) forms of UO2+x, which is also observed after the Ar+ etching of samples AP7 (Ar+) (Table 2). These data show that within the measurement error uranium quantitatively has an ionic composition weakly but depends on the crystallographic orientation. Core Electron XPS Range. Hydrocarbons and water molecules that contain oxygen adsorb on the surface of the samples during the handling. The surfaces of the studied samples, as already noted, were not cleaned with Ar+ ions in order not to destroy its initial structure. The O 1s spectrum of the studied samples consists of a relatively sharp peak at Eb(O 1s) = 530.1 eV, Γ(O 1s) = 1.0 eV (AP3) that is typical for crystalline UO2 (Figure 5, Table 3). At the higher BE side from the basic peak two low-intensity peaks at 531.1 and 531.9 eV of the surface oxygen were observed. The first peak can be attributed to the OH− group and the other one to the CO32− group. After 129Xe23+ irradiation the O 1s intensity grows significantly due to the substrate oxygen (see Table 2). This causes a significant change in UO2+x surface composition. The 238U31+ irradiations affect the XPS structure to a smaller extent. Results of the quantitative analysis on the basis of the U 4f and O 1s intensities show that the number of oxygen atoms in UO2+x is greater than 2.8. For example, for sample AP3 it is 3.6, and for OB3 it is 15. The corresponding kO values calculated on the basis of the U 5f intensity are 2.08 and 2.14 (Table 2). After the 129Xe23+ irradiation of samples OB1−4 the oxygen coefficient kO grows, U4+ content drops, and U5+ and U6+ content grows. After the 238U31+ irradiation of samples OB5−7 the oxygen coefficient kO decreases, U4+ content grows, and U5+ and U6+ content decreases. The penetration depth of xenon and uranium ions is ∼6.5 μm. Since the film thickness is ∼150 nm, the ions penetrate through the film and stop in the substrate, thus making it impossible to observe the xenon lines in the

the amount of oxygen remains constant on them within the experimental error (Table 2). Herewith, the value of kO increases in line AP5 (UO2 (001)), AP7 (UO2 (111)), and AP6 (UO2 (110)) for which the concentration of U4+ decreases and the concentration of U5+ and U6+ ions increases. 129 Xe23+ irradiation (92 MeV energy, 4.8 × 1015 ions/cm2 fluence, penetration depth ≈ 6.5 μm) of UO2 films (OB1−4) on the LSAT substrates was conducted in order to simulate the damage from nuclear fission fragments in nuclear fuel. The mass and energy of ions used for the irradiation are typical for fission fragments. The U 5f intensities, oxygen coefficients (kO), and ionic compositions (k, %) are given in Table 2 for each sample. In all cases the 129Xe23+ irradiation causes severe damage. This conclusion can be drawn from the fact that the structure disappears in the U 4f spectrum and the O 1s XPS peak grows significantly due to substrate oxygen in island formations with low uranium content on the substrate. This results in strong changes in UO2+x composition (Table 2). For example, the surface of unirradiated sample AP3 has the following elemental composition relative to the uranium atom: U1.00O3.6C1.8. The surface of complementary Xe-irradiated sample OB3 has the following composition: U1.00O15Ta3.9C167. Calculations of the elemental compositions involved only the O 1s intensity at 530.1 eV, the Ta 4f intensity, and the total C 1s intensity. The change in the surface composition can be explained by the fact that the single-crystalline film was severely damaged by Xe ions, and atoms of, for example, tantalum and oxygen from the substrate emerged to the surface. As a result, uranium oxide on the surface became amorphous. This led to a decrease of intensity and destruction of the U 4f spectrum, which is employed for quantitative ionic analysis on the basis of the U 4f BE. Since in the U 4f and U 5f BE ranges XPS peaks of other elements involved are absent, although the U 4f peak is of low intensity and not structured, the technique suggested in this work allows an evaluation of uranium ionic composition in the studied sample (Table 2), which make this technique original. The U 4f fine structure reflecting the OVMO structure vanishes since the chemical bond changes significantly (Figure 6b). In this case determination of the oxygen coefficient and ionic composition on the basis of the U 4f and O 1s intensities (traditional for XPS) is practically impossible. However, the technique based on the U 5f intensity, as shown in this work, allows one to evaluate the oxygen coefficient and ionic composition (Table 2). This is possible because the U 5f and U 4f photoionization cross sections are high, which means that the U 5f and U 4f peaks are intense, and the U 5f and U 4f BE ranges do not contain peaks of other elements.81 The 129Xe23+ irradiation results in a significant decrease of 4+ U content and increase of U5+ and U6+ content (Table 2). It is especially noticeable for OB1, where U4+ ions are practically absent. Since the solubility of uranium ions strongly depends on the oxidation state (U6+ more soluble than U4+ by several orders of magnitude), the absence of U4+ in sample OB1 is supported by a higher content (∼by a factor of 100) of dissolved uranium ions in deionized water as compared to samples OB2−4. For surfaces of the unirradiated films on the YSZ substrates (AP5−7) a significant decrease of U4+ content and increase of U5+ and U6+ content compared to AP1−4 on the LSAT substrates was observed (Table 2). One of the reasons for this is the change in composition and structure of the lattice in the YSZ substrates compared to those in the LSAT substrates that 8066

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry spectra. The xenon ions destroy the lattice, but since the fluence of uranium ions is low, uranium ions only rearrange the oxygen lattice that results in a more stable structure. Despite the presence of hydrocarbons and oxygen-containing compounds, uranium dioxide core−electron XPS peaks are observed to be relatively intense and well resolved. Thus, the U 4f XPS spectrum was observed as an intense spin−orbit split (ΔEsl = 10.8 eV) doublet (Figures 6a and 7a, Table 3). This spectrum is structured and widened due to the presence of uranium ions of different oxidation states and contains up to 30% of intensity from the shake-up satellites. The presence of the shake-up satellites indicates the long-range ordering in the films. Destruction of the long-range order leads to vanishing of the structure and satellites (Figure 6b). Since the U 4f spectrum can contain the shake-up satellites at ∼7 (U4+), ∼8 (U5+), and ∼4 and ∼10 eV (U6+),50,60 this spectrum is complicated and structured (Figures 6a and 7a). Having suggested that the samples contain three types of uranium ions (U4+, U5+, U6+) and calculated their quantitative compositions on the basis of the U 5f relative intensity, the U 4f spectra were decomposed into components. The criteria of accuracy of this decomposition were the difference of the total area of the calculated and the experimental peaks and the best match with the data obtained on the basis of the U 5f intensity (Figures 6 and 7, Table 2). Poor resolution of the XPS spectra can lead to a high error. Results of this decomposition are given in parentheses in Table 2. In some cases they differ significantly from the data obtained on the basis of the U 5f intensity. Parameters of the peaks and satellites for uranium ions are given in Table 3. The peaks of uranium ions of different oxidation states were suggested to have the same fwhms. The smallest fwhm was Γ(U 4f7/2) = 1.4 eV. The narrowest U 4f peak was expected from the U(5f0)6+ ions that do not have the U 5f electrons responsible for the possible widening due to the multiplet splitting. Indeed, for reference samples of PbUO4 and BaUO4 the U 4f7/2 peaks are 1.1 and 1.2 eV wide, respectively. Thus, having neglected the U 4f7/2 multiplet splitting, one can expect the U 4f7/2 peak of the monophase UO2 (CaF2) single crystal to be 1.1 eV wide. Therefore, the total fwhm of the U 4f7/2 peak Γ(U 4f7/2) measured relative to Γ(C 1s = 1.3 eV) is an important qualitative characteristic (Table 3). Depending on the concentrations of included uranium ions in the studied films, the shape of the U 4f peak will change. Qualitative correlation of the total U 4f7/2 fwhm with concentrations of uranium ions of different oxidation states has to be observed. Thus, for OB1,4 films U4+ concentrations are significantly lower and the Γ(U 4f7/2) are also lower compared to those of AP1,4 films (Tables 2 and 3). The C 1s spectrum of the studied samples consists of the basic peak at E b = 285.0 eV attributed to saturated hydrocarbons, the peak at 286.5 eV attributed to carbon connected with oxygen, and the peak at 288.7 eV associated with the CO32− group on the surface (Figure 8). After 129Xe23+ irradiation the C 1s intensities grow; 238U31+ irradiation practically does not cause any changes in the C 1s intensity. The error in the quantitative elemental and ionic analysis of the studied samples grows because the XPS spectra of the core levels exhibit extra structure due to the multiplet splitting and secondary electronic processes (many-body perturbation and dynamic effect). Since the many-body perturbation results in shake-up satellites at the higher BE side from the main peaks, satellite intensities can be partially taken into account. The

Figure 8. XPS narrow scan of C 1s from a UO2 film (samples AP3).

dynamic effect can be hardly taken into account, but its probability is pretty low in the considered spectra. All of the above can increase the error in the elemental and ionic analysis by about ±5%. This work evaluated the oxygen coefficient for oxides UO2+x on the basis of the U 4f7/2 and O 1s intensities (Table 2 and Figures 5 and 6). As mentioned above, oxygen adsorption at the surface makes practically impossible a correct evaluation of the oxygen coefficient for UO2+x (see Table 2). Since in the ratio nO/nU of oxygen, nO, and uranium, nU, the number of atoms changes in the range from 2.8 to 3.9, clusters with an excess of oxygen can form on the surface. Treatment with Ar+ ions of the surface of sample AP7 for 120 s leads to removal of the excessive oxygen, and the value (UO1.98), found based on the U 4f7/2 and O 1s intensities, is close to 2.0 ± 0.1. However, the value of kO found based on the U 5f relative intensity electrons reduced only to 2.11 and not to 2.00 as should be expected. This is related, possibly, to the fact that annealing of the sample did not take place to restore its UO2 (CaF2) lattice during presence in a vacuum of the chamber. The 129Xe23+ irradiation of the uranium oxide films damaged the surface significantly. The U 4f spectrum smears and the shake-up satellites vanish. This confirms the disappearance of the regular crystal structure in the samples (Figure 6b). In this case, the ionic analysis of the film’s surface is practically impossible on the basis of the U 4f and O 1s parameters. A significant decrease of the U 5f intensity takes place due to the change in the surface ionic composition. Despite severe damage of the surface, it is possible to evaluate the ionic composition on the basis of the U 5f intensity. Thus, the U4+ content drops while the U5+ and U6+ content grows significantly (see Table 2). On the basis of these data one can conclude that 129Xe23+ irradiation leads to both long-range order destruction and short-range order destruction of the uranium close environment, which results in an increase of uranium oxidation state and restructuring of oxygen ions in the uranium environment. The 238U31+ irradiations lead to a partial long-range order destruction and formation of lattice defects. X-ray diffraction study showed that the ion irradiations resulted in diffraction peak broadening due to a decrease in the size of coherent scattering domains that is consistent with destruction of the long-range order. OKLL Auger Spectral Structure of Uranium Dioxide Film (AP3). The valence band XPS of uranium dioxide film (AP3) and the calculation results were used for a qualitative explanation of the Auger O KLL spectra structure from this film (Figures 3 and 8). 8067

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry

depend on the nature of the substrate and weakly on the crystallographic orientation. On the basis of the spectral parameters it was established that uranium dioxide films AP2,3 on the LSAT substrates have closest to stoichiometric values of the kO, and from the XRD and EBDS results it follows that these films have a regular single-crystal structure. The XRD and EBSD results indicate that uranium oxide films on the YSZ substrates have singlecrystal structure; however, they have the highest oxygen coefficient kO. This difference is possibly related to the different nature of the substrates. The 129Xe23+ irradiation (92 MeV, fluence 4.8 × 1015 ions/ cm2) of uranium dioxide films on the LSAT substrates was shown to destroy both long-range ordering and uranium close environment, which results in an increase of uranium oxidation state and regrouping of oxygen ions in the uranium close environment. The 238U31+ irradiations (110 MeV, fluence 5 × 1010, 5 × 1011, and 5 × 1012 ions/cm2) of uranium dioxide films on the YSZ substrates was shown to form the lattice damage only with partial destruction of the long-range ordering. This is mainly due to low fluencies. However, the 238U31+ irradiations caused changes in the ionic composition of the samples: an increase in U 4+ concentration and a decrease in U 6+ concentration were observed; hence, the oxygen coefficient (kO) decreased compared to that for the unirradiated counterparts.

The O KLL Auger spectrum of Al2O3, where the O 2s shell participates weakly in the IVMO formation, consists of three well-resolved low structured ∼9 eV wide components reflecting the OKL2−3L2−3 (O 1s ← O 2p), OKL1L2−3 (O 1s ← O 2s,2p), and OKL1L1 (O 1s ← O 2s) electronic transitions.82 Relative intensities of these peaks given as the ratios of Auger peak intensities to the O 1s XPS peak intensity are important fundamental values. They allow a quantitative comparison of partial electronic densities, e.g., of the O 2p states on oxygen ions in different oxides.82 The O KLL Auger spectrum of uranium dioxide film (sample AP3) measured in this work consists of three structured bands (Figure 9). This structure is partially due to the oxygen-



AUTHOR INFORMATION

Corresponding Author

*Present address: Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EQ, United Kingdom. Phone: +44 1223 768357. E-mail: apopel@cantab. net.

Figure 9. Auger spectrum of O KLL from a UO2 film (sample AP3).

Notes

containing impurities on the sample surface. Despite this fact, a qualitative interpretation of this spectrum is possible. In the O KLL Auger spectrum of AP3 the OKL2−3L2−3 (O 1s ← O 2p) peak at 973.8 eV and Γ ≈ 3.6 eV reflects the density of the filled O 2p states (Figure 9). The fwhm of this peak qualitatively agrees with the sum of the XPS O 1s fwhm (Γ = 1.0 eV, Figure 5) and the O 2p band fwhm (Γ ≈ 4.4 eV, Figure 3). The OKL1L2−3 (O 1s ← O 2s,2p) band reflecting the O 2p and O 2s densities of states (DOS) is structured. The OKL1L1 (O 1s ← O 2s) band reflecting the O 2s DOS is also structured due to participation of the O 2s electrons in the IVMO formation. This band is ∼15 eV wide, which is comparable with the IVMO XPS band fwhm. This confirms the participation of the O 2s AOs in the IVMO formation (Figures 3 and 9). These results agree with the Auger O KLL results for UO2 (ref 83).

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The irradiation experiment was performed at the Grand Accélérateur National d’Ions Lourds (GANIL) Caen, France, and supported by the French Network EMIR. The support in planning and execution of the experiment by the CIMAPCIRIL and the GANIL staff, especially I. Monnet, C. Grygiel, T. Madi, and F. Durantel, is much appreciated. The work was supported by RFBR grant no. 16-03-00914-a and partially supported by M.V. Lomonosov Moscow State University Program of Development. A.J.P. acknowledges funding from the UK EPSRC (grant EP/I036400/1) and Radioactive Waste Management Ltd. (formerly the Radioactive Waste Management Directorate of the UK Nuclear Decommissioning Authority, contract NPO004411A-EPS02), a maintenance grant from the Russian Foundation for Basic Research (projects 13-03-90916) and CSAR bursary. Thanks are given to A.M. Adamska, G.I. Lampronti, V.A. Lebedev, P.G. Martin, L. Payne, and A.A. Shiryaev for their help in characterization of the samples.



CONCLUSIONS XPS determination of the oxygen coefficient kO = 2 + x and ionic (U4+, U5+, and U6+) composition of oxides UO2+x formed on the surfaces of differently oriented (hkl) planes of thin UO2 films on the LSAT and YSZ substrates was performed. The U 4f and O 1s core−electron peak intensities as well as the U 5f relative intensity before and after 129Xe23+ (92 MeV, fluence 4.8 × 1015 ions/cm2) and 238U31+ (110 MeV, fluence 5 × 1010, 5 × 1011, and 5 × 1012 ions/cm2) irradiations were employed. It was found that the presence of uranium dioxide film in air results in formation of oxide UO2+x on the surface with mean oxygen coefficients kO from 2.07 to 2.11 on the LSAT and from 2.17 to 2.23 on the YSZ substrates. These oxygen coefficients



REFERENCES

(1) He, H.; Qin, Z.; Shoesmith, D. W. Electrochim. Acta 2010, 56, 53−60. (2) Matzke, Hj. J. Nucl. Mater. 1992, 190, 101−106. (3) Sonoda, T.; Kinoshita, M.; Ishikawa, N.; Sataka, M.; Iwase, A.; Yasunaga, K. Nucl. Instrum. Methods Phys. Res., Sect. B 2010, 268, 3277−3281.

8068

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry (4) Matzke, Hj.; Lucuta, P. G.; Wiss, T. Nucl. Instrum. Methods Phys. Res., Sect. B 2000, 166−167, 920−926. (5) Wiss, T.; Matzke, Hj.; Trautmann, C.; Toulemonde, M.; Klaumünzer, S. Nucl. Instrum. Methods Phys. Res., Sect. B 1997, 122, 583−588. (6) Opel, K.; Weiß, S.; Hübener, S.; Zänker, H.; Bernhard, G. Radiochim. Acta 2007, 95, 143−149. (7) He, H.; Zhu, R. K.; Qin, Z.; Keech, P.; Ding, Z.; Shoesmith, D. W. J. Electrochem. Soc. 2009, 156, C87−C94. (8) Shoesmith, D. W.; Sunder, S. J. Nucl. Mater. 1992, 190, 20−35. (9) Andersson, D. A.; Uberuaga, B. P.; Nerikar, P. V.; Unal, C.; Stanek, C. R. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 84, 054105. (10) Andersson, D. A.; Lezama, J.; Uberuaga, B. P.; Deo, C.; Conradson, S. D. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 79, 024110. (11) Crocombette, J.-P.; Torumba, D.; Chartier, A. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 83, 184107. (12) Wang, J.; Ewing, R. C.; Becker, U. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 88, 024109. (13) Geng, H. Y.; Chen, Y.; Kaneta, Y.; Kinoshita, M. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 75, 054111. (14) Rak, Zs.; Ewing, R. C.; Becker, U. Surf. Sci. 2013, 608, 180−187. (15) Wang, D.; van Gunsteren, W. F.; Chai, Z. Chem. Soc. Rev. 2012, 41, 5836−5865. (16) Boettger, J. C.; Ray, A. K. Int. J. Quantum Chem. 2002, 90, 1470−1477. (17) Fuggle, J. C.; Mårtensson, N. J. J. Electron Spectrosc. Relat. Phenom. 1980, 21, 275−281. (18) Moser, H. R.; Delley, B.; Schneider, W. D.; Baer, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 1984, 29, 2947. (19) Fu, X.; Wang, X.; Zhao, Z.; Yu, Y. Surf. Rev. Lett. 2003, 10, 381− 386. (20) Moore, K. T.; van der Laan, G. Rev. Mod. Phys. 2009, 81, 235. (21) Opeil, C. P.; Schulze, R. K.; Manley, M. E.; Lashley, J. C.; Hults, W. L.; Hanrahan, R. J., Jr.; Smith, J. L.; Mihaila, B.; Blagoev, K. B.; Albers, R. C.; Littlewood, P. B. Phys. Rev. B: Condens. Matter Mater. Phys. 2006, 73, 165109. (22) Kudin, K. N.; Scuseria, G. E.; Martin, R. L. Phys. Rev. Lett. 2002, 89, 266402. (23) Prodan, I. D.; Scuseria, G. E.; Martin, R. L. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 76, 033101. (24) Prodan, I. D.; Scuseria, G. E.; Martin, R. L. Phys. Rev. B: Condens. Matter Mater. Phys. 2006, 73, 045104. (25) Dorado, B.; Amadon, B.; Freyss, M.; Bertolus, M. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 79, 235125. (26) Gubanov, V. A.; Rosen, A.; Ellis, D. E. J. Phys. Chem. Solids 1979, 40, 17−28. (27) Suzuki, C.; Nishi, T.; Nakada, M.; Akabori, M.; Hirata, M.; Kaji, Y. Int. J. Quantum Chem. 2009, 109, 2744−2752. (28) Wen, X.-D.; Martin, R. L.; Roy, L. E.; Scuseria, G. E.; Rudin, S. P.; Batista, E. R.; McCleskey, T. M.; Scott, B. L.; Bauer, E.; Joyce, J. J.; Durakiewicz, T. J. Chem. Phys. 2012, 137, 154707. (29) Wen, X.-D.; Martin, R. L.; Henderson, T. M.; Scuseria, G. E. Chem. Rev. 2013, 113, 1063−1096. (30) Suzuki, C.; Nishi, T.; Nakada, M.; Tsuru, T.; Akabori, M.; Hirata, M.; Kaji, Y. J. Phys. Chem. Solids 2013, 74, 1769−1774. (31) Yang, Y.; Zhang, P. J. Appl. Phys. 2013, 113, 013501. (32) Yun, Y.; Rusz, J.; Suzuki, M.-T.; Oppeneer, P. M. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 83, 075109. (33) Shi, H.; Chu, M.; Zhang, P. J. Nucl. Mater. 2010, 400, 151−156. (34) Petit, L.; Svane, A.; Szotek, Z.; Temmerman, W. M.; Stocks, G. M. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 045108. (35) Naegele, J. R.; Ghijsen, J.; Manes, L. Struct. Bonding (Berlin) 1985, 59/60, 197−262. (36) Seibert, A.; Gouder, T.; Huber, F. Radiochim. Acta 2009, 97, 247−250. (37) Gouder, T.; Seibert, A.; Havela, L.; Rebizant, J. Surf. Sci. 2007, 601, L77−L80.

(38) Idriss, H. Surf. Sci. Rep. 2010, 65, 67−109. (39) Miserque, F.; Gouder, T.; Wegen, D. H.; Bottomley, P. D. W. J. Nucl. Mater. 2001, 298, 280−290. (40) Burrell, A. K.; McCleskey, T. M.; Shukla, P.; Wang, H.; Durakiewicz, T.; Moore, D. P.; Olson, C. G.; Joyce, J. J.; Jia, Q. Adv. Mater. 2007, 19, 3559−3563. (41) Beatham, N.; Orchard, A. F.; Thornton, G. J. Electron Spectrosc. Relat. Phenom. 1980, 19, 205−211. (42) Teterin, Yu. A.; Maslakov, K. I.; Ryzhkov, M. V.; Traparic, O. A.; Vukvcevic, L.; Teterin, A.; Yu; Panov, A. D. Radiochemistry 2005, 47, 215−224. (43) Chadwick, D.; Graham, J. Nature, Phys. Sci. 1972, 237, 127−128. (44) Teterin, Yu. A.; Kulakov, V. M.; Baev, A. S.; Zelenkov, A. G.; Nevzorov, N. B.; Melnikov, I. V.; Sreltsov, V. A.; Mashirov, L. G.; Suglobov, D. N. Dokl. Akad. Nauk SSSR (Rep. Acad. Sci. USSR) 1980, 255, 434−437 (in Russian). (45) Teterin, Yu. A.; Kulakov, V. M.; Baev, A. S.; Nevzorov, N. B.; Melnikov, I. V.; Streltsov, V. A.; Mashirov, L. G.; Suglobov, D. N.; Zelenkov, A. G. Phys. Chem. Miner. 1981, 7, 151−158. (46) Miyake, C.; Sakurai, H.; Imoto, S. Chem. Phys. Lett. 1975, 36, 158−160. (47) Allen, G. C.; Holmes, N. R. Can. J. Appl. Spectrosc. 1993, 38, 124−130. (48) Van den Berghe, S.; Miserque, F.; Gouder, T.; Gaudreau, B.; Verwerft, M. J. Nucl. Mater. 2001, 294, 168−174. (49) Teterin, A. Yu.; Teterin, Yu. A.; Maslakov, K. I.; Batuk, O. N.; Kalmykov, S. N.; Zakharova, E. V. Radiochemistry 2009, 51, 450−457. (50) Teterin, Yu. A.; Teterin, A. Yu. Russ. Chem. Rev. 2004, 73, 541− 580. (51) Ulrich, K.-U.; Ilton, E. S.; Veeramani, H.; Sharp, J. O.; BernierLatmani, R.; Schofield, E. J.; Bargar, J. R.; Giammar, D. E. Geochim. Cosmochim. Acta 2009, 73, 6065−6083. (52) Veal, B. W.; Lam, D. J. Phys. Rev. B 1974, 10, 4902. (53) Veal, B. W.; Diamond, H.; Hoekstra, H. R. Phys. Rev. B 1977, 15, 2929. (54) Veal, B. W.; Lam, D. J.; Diamond, H. Physica B+C 1977, 86−88, 1193−1194. (55) Cox, L. E.; Farr, J. D. Phys. Rev. B: Condens. Matter Mater. Phys. 1989, 39, 11142. (56) Gouder, T.; Colmenares, C.; Naegele, J. R.; Verbist, J. Surf. Sci. 1990, 235, 280−286. (57) Stumpf, S.; Seibert, A.; Gouder, T.; Huber, F.; Wiss, T.; Römer, J. J. Nucl. Mater. 2009, 385, 208−211. (58) Chen, Q.; Lai, X.; Bai, B.; Chu, M. Appl. Surf. Sci. 2010, 256, 3047−3050. (59) Teterin, Yu. A.; Baev, A. S. Rentgenovskaya Fotoelectronnaya Spectroscopiya Soedinenii Legkikh Aktinoidov (X-Ray Photoelectron Spectroscopy of Light Actinide Compounds); TsNIIatominform: Moscow, 1986; p 104 (in Russian). (60) Ilton, E. S.; Bagus, P. S. Surf. Interface Anal. 2011, 43, 1549− 1560. (61) Teterin, A. Yu.; Ryzhkov, M. V.; Teterin, Yu. A.; Panov, A. D.; Nikitin, A. C.; Ivanov, K. E.; Utkin, I. O. Radiochemistry 2002, 44, 224−233. (62) Ilton, E. S.; Haiduc, A.; Cahill, C. L.; Felmy, A. R. Inorg. Chem. 2005, 44, 2986−2988. (63) Ilton, E. S.; Bagus, P. S. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 71, 195121. (64) Pireaux, J. J.; Riga, J.; Thibaut, E.; Tenret-Noel, C.; Caudano, R.; Verbist, J. J. Chem. Phys. 1977, 22, 113−120. (65) Yarzhemsky, V. G.; Nefedov, V. I.; Trzhaskovskaya, M. B.; Band, I. M.; Teterin, Yu. A.; Teterin, A. Yu. J. Surf. Invest.: X-ray, Synchrotron Neutron Technol. 2001, 7, 32 (in Russian). (66) Yarzhemsky, V. G.; Nefedov, V. I.; Trzhaskovskaya, M. B.; Band, I. M.; Szargan, R. J. Electron Spectrosc. Relat. Phenom. 2002, 123, 1−10. (67) Teterin, Yu. A.; Ivanov, K. E.; Teterin, A.; Yu; Lebedev, A. M.; Utkin, I. O.; Vukchevich, L. J. Electron Spectrosc. Relat. Phenom. 1999, 101−103, 401−405. 8069

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070

Article

Inorganic Chemistry (68) Teterin, A. Yu.; Ryzhkov, M. V.; Teterin, Yu. A.; Maslakov, K. I.; Reich, T.; Molodtsov, S. L. J. Struct. Chem. 2011, 52, 295−303. (69) Teterin, Yu. A.; Teterin, A. Yu. Nucl. Technol. Radiat. Prot. 2004, 19, 3−14. (70) Tobin, J. G.; Yu, S.-W. Phys. Rev. Lett. 2011, 107, 167406. (71) Bao, Z.; Springell, R.; Walker, H. C.; Leiste, H.; Kuebel, K.; Prang, R.; Nisbet, G.; Langridge, S.; Ward, R. C. C.; Gouder, T.; Caciuffo, R.; Lander, G. H. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 88, 134426. (72) Chakoumakos, B. C.; Schlom, D. G.; Urbanik, M.; Luine, J. J. Appl. Phys. 1998, 83, 1979−1982. (73) Strehle, M. M.; Heuser, B. J.; Elbakhshwan, M. S.; Han, X.; Gennardo, D. J.; Pappas, H. K.; Ju, H. Thin Solid Films 2012, 520, 5616−5626. (74) Ziegler, J. F.; Biersack, J. P.; Ziegler, M. D. The Stopping and Range of Ions in Matter; SRIM Co.: Chester, MD, 2008. (75) Shirley, D. A. Phys. Rev. B 1972, 5, 4709. (76) Gouder, T.; Havela, L. Microchim. Acta 2002, 138, 207−215. (77) Frazer, B. C.; Shirane, G.; Cox, D. E.; Olsen, C. E. Phys. Rev. 1965, 140, A1448. (78) Rafalsky, R. P.; Alexeev, V. A.; Ananyeva, L. A. Geokhimiya (Geochemistry) 1979, 11, 1601 (in Russian). (79) Stubbs, J. E.; Chaka, A. M.; Ilton, E. S.; Biwer, C. A.; Engelhard, M. H.; Bargar, J. R.; Eng, P. J. Phys. Rev. Lett. 2015, 114, 246103. (80) Il’in, E. G.; Parshakov, A. S.; Teterin, A. Yu.; Maslakov, K. I.; Teterin, Yu. A. Russ. J. Inorg. Chem. 2011, 56, 1788−1793. (81) Scofield, J. H. J. Electron Spectrosc. Relat. Phenom. 1976, 8, 129− 137. (82) Teterin, Y. A.; Ivanov, K. E.; Teterin, A. Y.; Lebedev, A. M.; Utkin, I. O.; Vukchevich, L. J. Electron Spectrosc. Relat. Phenom. 1999, 101−103, 401−405. (83) Teterin, Yu. A.; Teterin, A. Yu. Radiochemistry 2005, 47, 440− 446.

8070

DOI: 10.1021/acs.inorgchem.6b01184 Inorg. Chem. 2016, 55, 8059−8070